Ep. 61 - Blind With Scissors: The NSA's MCP Warning for Every Agentic AI Deployment
The NSA just published a rare advisory on the Model Context Protocol (MCP)—the plumbing under nearly every agentic AI deployment of the last 18 months—and the verdict is stark: optional authentication, no token lifecycle, silent behavior changes, and no logging to catch any of it. Host Tova Dvorin sits down with defensive cybersecurity expert Adrian Culley to unpack the eight risk categories, the WhatsApp and GitHub MCP exploits, and why MCP is now a testable validation surface.
Key takeaways:
00:00 Introduction to MCP and its significance
00:28 NSA's advisory on MCP vulnerabilities
01:14 What is the Model Context Protocol?
02:01 Security challenges of MCP
03:04 The inversion of client-server architecture
04:22 Access control issues with MCP
05:25 The multi-server problem
06:10 WhatsApp case study
07:10 Poor approval workflows in MCP
08:03 GitHub MCP access issues
09:38 Serialization and context injection risks
10:32 Token and session security concerns
12:22 Inconsistent behavior and audit logging
13:26 Recommendations from the NSA advisory
Subscribe for weekly insights on cybersecurity and AI developments.